Understand Singular Value Decomposition (SVD) Intuitively
Authors: Bill Cockerill
Working with Shiny < 1 year
Abstract: Singular Value Decomposition (SVD) is a fundamental data analysis technique used in key Data Science methods such as NLP and Machine Learning. Play with this app to help you understand how SVD works intuitively.
Full Description: Singular Value Decomposition (SVD) is a powerful data analysis technique with many uses. It includes data dimension reduction prior to Machine Learning (using Principal Components Analysis) and solving linear equations. As well as image compression and improving our ability to search text through Latent Semantic Analysis (LSA), both of which you can experiment with using this app.
The betterexplained website, the Feynman Technique, and David Robinson's empirical Bayesian methods are all inspirations for explaining this important technique intuitively. In the spirit of those teachers, this app tries not to assume any previous technical knowledge.
By playing with image compression and LSA with text in this app I hope you can build a more intuitive understanding of SVD. And from this greater intuition, you might use SVD more confidently and appropriately in your data analysis.
Category: Education
Keywords: SVD,image compression,PCA,LSA,education,intuition
Shiny app: https://bill-c.shinyapps.io/SVD_intuition/
Repo: GitHub - billster45/SVD_intuition_app: View the shiny app deployed here:
RStudio Cloud: Posit Cloud
Thumbnail:
Full image:
